Dear All,
I am a civil engineer, specialized in water resources engineering.
My aim is to analyze 30 years of daily stream flow data in an Excel table with about 300 rows x 31 columns.
The data table contains negative values and strings which I have to discard first.
Then I want to create a column vector of data for further analysis.
The Smath program shown below gives the result, but takes a long time (about 50 seconds- see image).
As I build up the program for further analysis, the time grows up to even 5 minutes or more.
However, my Mathcad 15 program yields the result almost immediately.

Can anyone help to resolve this please? Thanks in advance.
ndtma

---

DailyQ_Malwala.xls (121 KiB) downloaded 87 time(s).

Hello ndtma

I attach my proposal (approx. 25s on my notebook). The first part takes advantage from how are built matrices in SMath. The slower part is the 2nd (as for now I don't have ideas on how to improve this part).

[edit] please note that the first part is not strictly required, anyway seems to me (?) that there's a small speed improvement using a vector instead of the linear indexing on the matrix

---

Hello Davide

Thanks for your valuable proposal, and I am awaiting more improvements.

Best Regards
ndtma

12s average (not immediate, however faster than the previous solution)

---

Kalu Ganga-Malwala Operation Study-Simple-R1 (1).sm (44 KiB) downloaded 106 time(s).

Hello Davide

Thanks for the second improvement.

Best Regards
ndtma

I was faced with a similar problem of a slow function (in fact EXCEL interop call). I could not speed up the function so I have created if() loops that will run the function only when input has changed. This makes sure that when workbook automatically recalculates it will not get "stuck" on a slow function unless necessary.

A sample is attached. In the workbook c:a+b is an example function, while a+b is input.

Hope that helps

---

input_CHK.sm (5 KiB) downloaded 100 time(s).